Ontology based customer support techniques

Abstract

Methods and systems for providing customer support in response to support communication are disclosed. Such communications may be structured or unstructured, and unstructured communications may be subject to further processing. A determination is made at least as to one or more relevant support classes, as well as to whether further support action(s) should be initiated. When determined that at least a further support action should be initiated, aspects of the support action may be determined and initiated.

What is claimed is: 
     
       1. A computer-implemented method on a computer system of a resource provider, the computer-implemented method comprising:
 generating a tree data structure associated with computing services of the resource provider, nodes of the tree data structure indicative of the computing services and associated with support classes of the resource provider, the computing services remotely offered to a plurality of clients based at least in part on computing resource configurations of virtual instances hosted on the computer system; 
 receiving, from a public communication network, an unstructured communication of a client of the resource provider, the unstructured communication related to a computing service of the resource provider; 
 generating, based at least in part on natural language processing (NLP) of the unstructured communication according to a format for a traversal of the tree data structure, a structured communication indicating a computing issue associated with the computing service, the structured communication having the format for the traversal, the computing service corresponding to a virtual instance of the virtual instances; 
 inferring, from the structured communication, client information indicating a potential computing resource configuration of the resource provider and associated with the computing service, the inferring comprising:
 comparing client data from the structured communication to resource provider data from the nodes of the tree data structure; 
 identifying a node of the tree data structure as representing a potential computing service provided to the client based at least in part on an outcome of the comparing; and 
 identifying the potential computing resource configuration based at least in part on the potential computing service; 
 
 determining, based at least in part on the client information and the structured communication, that the potential computing resource configuration comprises a computing resource configuration of the virtual instance, the computing resource configuration provided to the client from the computing resource configurations; 
 determining a support class relevant to the structured communication based at least in part on the computing resource configuration and the node representing the potential computing service; and 
 instructing the computing service about a set of actions associated with resolving the computing issue, the set of actions defined in a support workflow relevant to the support class, the computing service updating the virtual instance to resolve the computing issue based at least in part on the set of actions. 
 
     
     
       2. The computer-implemented method of  claim 1 , wherein generating the structured communication comprises:
 detecting a trigger indicating the unstructured communication between the client and a third party; and 
 receiving the unstructured communication based at least in part on the trigger; 
 wherein the structured communication is generated from the unstructured communication further based at least in part on content of the unstructured communication and the format. 
 
     
     
       3. The computer-implemented method of  claim 1 , further comprising:
 determining an appropriate workflow sequence for the support workflow based at least in part on the structured communication; 
 executing the determined appropriate workflow sequence; and 
 notifying the client of an execution of the workflow sequence. 
 
     
     
       4. The computer-implemented method of  claim 1 , wherein the client information and the structure communication analyzed to set the potential computing resource configuration as the computing resource configuration comprise at least one of: a characteristic of the client, a characteristic of the potential computing service, or content of the structured communication. 
     
     
       5. The computer-implemented method of  claim 1 , wherein analyzing the client information and the structured communication to set the potential computing resource configuration as the computing resource configuration comprises:
 determining that the structured communication comprises missing information to set the node as the node corresponding to the computing service; and 
 traversing, based at least in part on the missing information, the tree data structure to detect that the node is the only node that represents the computing service. 
 
     
     
       6. The computer-implemented method of  claim 1 , wherein analyzing the client information and the structured communication to set the potential computing resource configuration as the computing resource configuration comprises:
 determining that the structured communication comprises missing information to set the node as the node corresponding to the computing service; 
 requesting the missing information from the client; 
 receiving the missing information from the client; and 
 detecting that the node represents the computing service based at least in part on the mission information. 
 
     
     
       7. The computer-implemented method of  claim 1 , wherein analyzing the client information and the structured communication to set the potential computing resource configuration as the computing resource configuration comprises:
 determining a relationship between words of the structured communication, the relationship comprising at least one of: weights of the words or proximities between the words; and 
 setting the potential computing resource configuration as the computing resource configuration based at least in part on the relationship. 
 
     
     
       8. The computer-implemented method of  claim 1 , wherein the node is a child node of a parent node of the tree data structure, wherein the child node comprises a common characteristic of the potential computing service shared with the parent node and a particular characteristic of the potential computing service specific to the child node, and wherein identifying the node comprises selecting the node based at least in part on the client data and the structured communication comprising at least one of: the common characteristic or the particular characteristic. 
     
     
       9. A computer-implemented method on a computer system comprising:
 accessing an unstructured communication of a client of a resource provider, the unstructured communication about a computing issue associated with a computing service of the resource provider, the unstructured communication received from a public communication network, the computing service corresponding to a virtual instance; 
 accessing a tree associated with computing services of the resource provider, nodes of the tree indicative of the computing services, the computing services remotely offered to a plurality of clients based at least in part on computing resource configurations of the resource provider; 
 generating a communication from the unstructured communication based at least in part on natural language processing (NLP) of the unstructured communication according to a format for a traversal of the tree, the communication having the format; 
 inferring, based at least in part on the communication of the client, information indicating a potential computing resource configuration of the resource provider and associated with the computing service, the inferring comprising:
 identifying a node of the tree as representing a potential computing service provided to the client based at least in part on a comparison of first data from the communication to second data from the tree; and 
 identifying the potential computing resource configuration based at least in part on the potential computing service; 
 
 determining, based at least in part on the information that is inferred and the communication that is accessed, that the potential computing resource configuration comprises a computing resource configuration of the virtual instance, the computing resource configuration provided to the client from the computing resource configurations; and 
 instructing the computing service about a support action for the client and relevant to a resolution of the computing issue based at least in part on the computing resource configuration and the node representing the potential computing service. 
 
     
     
       10. The computer-implemented method of  claim 9 , wherein the support action is a part of a support workflow, and further comprising:
 determining a support class relevant to the communication based at least in part on the computing resource configuration and the node representing the potential computing service; 
 initiating the support workflow relevant to the support class; 
 tracking metrics of the support workflow; and 
 submitting workflow information related to the support workflow to an entity associated with the support class based at least in part on the metrics. 
 
     
     
       11. The computer-implemented method of  claim 10 , wherein the support class is determined from among a plurality of support classes organized in the tree based at least in part on the information that is inferred and the communication that is accessed. 
     
     
       12. The computer-implemented method of  claim 9 , wherein the information that is inferred comprises additional information not included in the communication that is accessed. 
     
     
       13. The computer-implemented method of  claim 9 , further comprising: tracking outcomes of past determination to initiate the support action for a plurality of clients, and wherein the support action is initiated based at least in part on the outcomes and a characteristic of the client. 
     
     
       14. The computer-implemented method of  claim 9 , wherein the communication comprises a structured communication, and wherein generating the communication further comprises:
 applying, to the unstructured communication, one or more of linguistic stemming, determination of word weights, or determination of word proximities to generate the structured communication according to a format for traversing the tree. 
 
     
     
       15. A system comprising:
 one or more processors; 
 memory comprising instructions that, upon execution with the one or more processors, cause the system to perform operations comprising:
 obtaining an unstructured communication of a client of a resource provider, the unstructured communication about a computing issue associated with a computing service of the resource provider and received from a public communication network, the computing service corresponding to a virtual instance; 
 accessing a tree associated with computing services of the resource provider, nodes of the tree indicative of the computing services and associated with support classes of the resource provider, the computing services remotely offered to a plurality of clients based at least in part on computing resource configurations of the resource provider; 
 generating a communication from the unstructured communication based at least in part on a natural language processing (NLP) of the unstructured communication according to a format for a traversal of the tree, the communication having the format; 
 inferring, based at least in part on the communication of the client, information indicating a potential computing resource configuration of the resource provider and associated with the computing service, the inferring comprising:
 identifying a node of the tree as representing a potential computing service provided to the client based at least in part on a comparison of first data from the communication to second data from the tree; and 
 identifying the potential computing resource configuration based at least in part on the potential computing service; 
 
 determining, based at least in part on the information and the communication, that the potential computing resource configuration comprises a computing resource configuration of the virtual instance, the computing resource configuration provided to the client from the computing resource configurations; 
 selecting a support class relevant to the communication from the support classes based at least in part on the computing resource configuration and the node representing the potential computing service; and 
 instructing the computing service about a set of actions associated with resolving the computing issue, the set of actions defined based at least in part on the support class. 
 
 
     
     
       16. The system of  claim 15 , wherein obtaining the unstructured communication of the client is based at least in part on a trigger, wherein the trigger is detected by a listener of the system, the listened is configured to track use on public communications network presences of handles or metadata tags, the handles comprising at least identities associated with the client, the metadata tags comprising at least hashtags. 
     
     
       17. The system of  claim 16 , wherein the trigger is detected use of at least one metadata tag among a plurality of predetermined triggering metadata tags. 
     
     
       18. The system of  claim 15 , wherein the communication is generated, by an ingestion engine of the system, as a structured communication from an unstructured communication of the client, the ingestion engine configured to generate the structured communication based at least in part on the format. 
     
     
       19. The system of  claim 18 , wherein accessing the tree, inferring the information, analyzing the information, and selecting the support class are implemented by a decision engine of the system, wherein the decision engine is configured to identify the support class to a workflow engine of the system. 
     
     
       20. The system of  claim 19 , wherein the workflow engine is configured to initiate a workflow corresponding to the support class and to track an outcome of the workflow, and wherein the decision engine is configured to utilize the outcome in response to a subsequent communication of the client to infer subsequent information from the subsequent communication.